Seminar in Finance
The History of Market Efficiency Concepts
Primeval Research on Changes in Stock Price
The work of Louis Bachelier
One of the most outstanding early works on the subject of market price behavior was the Doctoral
Dissertation of Louis Bachelier entitled “The Theory of Speculation” completed in 1900. Bachelier was a
student in the Department of Mathematics at the Sorbonne in Paris.
He began the study of the stock market assuming this was a source of time-series that would exhibit
discernable patterns. Very soon he began to notice that the price changes, which he had assumed would be
predictable, where simply random.
Bachileir’s work was never widely read during his life-time and in the ensuing decades. In fact, he had a
rather obscure career teaching in small, provincial French colleges. The work was “discovered” in 1954 by
Jimmie Savage and Paul Samuelson. Since then it has become something of a cult classic amongst those
studying securities markets for its high level of precision and academic rigor (for 1900).
Bacheleir also went pretty far down-the-road toward developing an option pricing model. His work had a
great influence on James Tobin in the 1960’s when he was working on the option pricing problem.
Alfred Cowles and the Cowles Commission
Alfred Cowles was a millionaire who lost a considerable fortune in the stock market crash of 1929. A few
days before The Crash William Peter Hamilton had published a now famed editorial in the Wall Street
Journal warning investors to liquidate their stock positions as disaster was eminent. After reading the
editorial (retrospectively) Cowles became fascinated with the question of whether movement of stocks and
stock markets were predictable. He gave an endowment to establish the Cowles Commission to begin
statistical studies of whether it was possible to predict such movements. The head of the project was
Professor Irving Fisher of Yale (who had also suffered a crippling personal loss in the Market Crash). The
Commission published its preliminary report in 1933 entitled “Can Stock Market Forecasters Forecast?”
The famed abstract consisted of three words, “It is doubtful.” The money Cowles gave to the foundation
still supports the publication of the prestigious academic journal Econometrica dedicated to empirical and
statistical analysis of economic phenomenon.
Other (Now) Famous Studies of Market Price Movements
In the two-and-a-half decades following the work of the Cowles Commission little work was done on
pricing phenomenon. Nobody really seemed very interested in the stock market at all during this period
(except, presumably, the traders working there themselves). During these lean years, two seminal works
were published in prestigious statistical journals. It is doubtful they were read at all by academics in
economics or business schools.
The first was by Holbrook Working in 1934. Working was actually looking at the prices of agricultural
commodities. He again noticed the random pattern in price changes. This and subsequent papers by
Working had an immense influence on later work on the pricing of futures contracts.
The second work was by Maurice Kendall a British statistician. He thought that stock and commodity
markets would provide a good source of statistical data to study that moved in some predictable pattern
(much like Bacheleir). Much to his surprise, he could find no such patterns in these prices.
Triumph of Efficient Market Orthodoxy in the 1960’s and 1970’s
The Great Papers of 1959
The year 1959 was a break through year in the study of price movements. Two papers were published that
formalized the basic framework of later studies of market efficiency. These two papers became the models
for empirical studies over the following decade, and their relevance is still considerable.
The first was published by M. F. M. Osborne an astrophysicist working at the US Naval Research
Laboratory. His paper was entitled “Brownian Motion in the Stock Market” published in Operations
Research. He found that the physicist’s model of random “Brownian Motion” did a wonderful job of
describing stock price movements. This model has been used since by academics in the field of Finance to
model stock prices. It is also often called the Random Walk.
The second paper was entitled “Stock Market ‘Patterns’ and Financial Analysts” published in the Journal
of Finance by Harry Roberts. The title implies that this was the first paper to take aim at the claims of
Technical Analysts that they could profit by studying patterns in stock price movements.
The Technicians Strike Back (But Then …..)
In a paper entitled “Price Movements in Speculative Markets” published in the Industrial Management
Review in 1961, Sidney Alexander took dead aim at the work of Harry Roberts. In this article Alexander
presented evidence that a great many technical strategies could be quite profitable. However, in a second
article published in 1964 he replicated his earlier study and found only minimal evidence of an profitable
Widening the Debate: Collecting all of the Evidence in One Place
The next seminal event in beginning a lively academic debate was the simple act of collecting all of the
available evidence together in one easily accessible volume. This role was filled by Paul Cootner with the
publication of the book The Random Character of Stock Market Prices in 1964. He brought together most
of the seminal works to date that suggested that stock price changes were simply random.
Defining the Efficient Market Hypotheses: The Work of Fama
The paper that refined and focused the debate into the form it has taken in Finance textbooks ever since was
Eugene Fama’s “Behavior of Stock Market Prices” published in the Journal of Business in 1965. This
article was essentially Fama’s entire PhD dissertation and ran 70 pages in the journal. Fama reviewed and
added some new tests to the previous attack by Harry Roberts on the Technical Analysts. He also began a
discussion of the value and useful life of new information coming into the market. You can see that this
laid the groundwork for the ensuing debate in terms of the three forms of efficiency in the widely used
Efficient Market Hypothesis (EMH).
Tests of the Semi-Strong Form of Market Efficiency
Early tests of the Semi-Strong Form of Market Efficiency were conducted by studying stock splits and
dividend announcements. Usually splits and dividends are perceived as good news (they also have
predictable impacts on stock price: in fact they cause them to fall). To a certain extent these actions are
reasonably easy to predict, even before the official announcement is made. In a paper published in 1969
Fama, Fisher, Jensen and Roll noticed that returns from stocks after splits and dividends are exactly the
same as for a random portfolio. What was interesting is that before the splits and dividends the stocks that
subsequently proclaimed the actions had higher returns that randomly chosen stocks. This means that the
information has already been available before the official announcement.
The second forum for testing the Semi-Strong Form was to look at the recommendations of brokerage
houses and the performance of mutual funds. The analysts at brokerages and fund managers are presumably
experts who have access to all the available information in the market. If they should be able to beat the
market if anyone can. However, most evidence collected through the 1970’s suggested that the stocks
recommended by brokerage houses do no better than randomly chosen stocks, and mutual funds seem to do
no better than a randomly chosen basket of stocks.
In all such studies it is vital to control for the risk of the investment. Anyone can get higher average returns
by investing in riskier stocks. The work on CAPM was very important in allowing researchers methods to
control for risk of portfolios.
Tests of the Strong Form of Market Efficiency
Clearly, it is difficult to conduct tests of the Strong Form of Market Efficiency mainly because insider
trading is illegal in most countries. However, at one time corporate managers were only required to report
sales or purchases of stock in their firms to the SEC after the transaction. Since the law has been amended
to require them to report transactions several months in advance. Using data from the SEC’s Insider Report
H. Kerr established that it was possible to make profits if you bought or sold stocks like the insiders if you
could carry out the transactions before the exchanges were reported in Insider Report but not after.
There is also an old Fundamentalist story that you should buy stocks with low P/E ratios (Price/Earnings).
However, official accounting data on earnings necessary to compute P/E ratios is only released quarterly.
Several studies have shown that if you could buy stocks that will subsequently report low P/E ratios (high
earnings) or sell stocks that will report low earnings, you could make handsome excess returns. But, if you
buy them on the day the earnings announcements become public, there is no return to be had.
The Faith Begins to Weaken: The 1980”s
Soon after the apparent triumph of the Efficient Markets orthodoxy, some troubling anomalies began to
crop-up in studies of stock market price changes. Apparent opportunities to make profits by following
specific buy-and-sell strategies through time were uncovered. Most troubling of all these patterns appear
The first discovered and most famous is the fabled January Effect. If you buy stocks on January 1 and sell
them on January 31, you seem to make superior returns (for the risk taken) than holding the stocks in the
other eleven months. This effect is usually blamed on fund managers dumping “losers” in their portfolio
before the end of the fiscal year.
Another is the Monday Effect. Again, if you buy first-thing on Monday morning and sell-out at the end of
the day superior profits can be realized beyond holding the other four days of the trading week.
More recently, as better data on intra-day trading has become available, an Opening Effect has also been
unveiled which suggests you can earn apparent excess returns by buying at the open and selling-out a few
A branch of macroeconomics has always been interested in the stock market as a factor that can influence
the economy. Probably driven by the experience of the 1929 Crash, there is a persistent belief that irrational
“Bubbles” can occur in markets. You are all aware that this argument has a powerful popular appeal.
Criticisms of the Market Efficiency story here attack the theory at its weakest point. Unfortunately, it is
pretty hard for either side to prove anything in this area. One of the things that has always bothered the
Macroeconomists is that stock prices are far more volatile than dividend streams, and this strikes them as
counter-intuitive. Some of the most interesting work in this area has been done by DeBondt and Thaler. In
some interesting theoretical papers they develop models where the underlying market is badly mis -priced
(an irrational bubble) but statistical tests such as those conducted by Fama in the 1960’s would still find no
time persistent patterns in (short-term) stock returns.
Evidence of Semi-Strong Inefficiency
Finally, over the last several decades, again with the accumulation of better and more accurate data,
researchers have begun to find more and more evidence of Fundamental Analysts whose advice would
seem to lead to considerable excess profit. The most famous of these is the Value Line Anomaly. If you
buy the s tocks rated 4 and 5 by the Value Line Investment Survey (immediately after the survey is updated)
investors appear to earn unwarranted returns.
Most of the material in this handout was gleaned from three sources. First, Peter L. Bernstein’s eminently
readable popular book Capital Ideas: The Improbable Origins of Modern Wall Street, second from Chapter
19 of the advanced textbook Portfolio and Investment Selection: Theory and Practice by Hiam Levy and
Marshall Sarnat, and finally from Eugene F. Fama’s article “Efficient Capital Markets: II” published in the
Journal of Finance in 1991.